{"id":138017,"date":"2026-08-13T00:03:28","date_gmt":"2026-08-13T00:03:28","guid":{"rendered":"https:\/\/www.europesays.com\/ai\/138017\/"},"modified":"2026-08-13T00:03:28","modified_gmt":"2026-08-13T00:03:28","slug":"researchers-observe-first-near-autonomous-ai-attack-on-government-target-in-taiwan","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ai\/138017\/","title":{"rendered":"Researchers observe first \u2018near-autonomous\u2019 AI attack on government target in Taiwan"},"content":{"rendered":"

Suspected Chinese hackers used open-source artificial intelligence models to run a cyberattack against the Taiwanese government in the first publicly known case of an autonomous AI hack hitting a government target, according to research published Wednesday.<\/p>\n

The hackers extracted more than 2,500 personnel records, among other data, in the \u201cnear-autonomous attack,\u201d researchers at Israeli cyber firm Dream wrote in a blog post<\/a>. The attackers set up the framework so that it could \u201cadapt mid-operation without human intervention.\u201d<\/p>\n

The framework \u201cimplements dedicated research phases it calls \u2018Learning Cycles\u2019 \u2014 autonomous sessions where the AI system searches vulnerability databases, GitHub repositories, and security research publications for techniques specifically applicable to its target government\u2019s infrastructure,\u201d the post reads.<\/p>\n

And then it kept going.<\/p>\n

\u201cThe attacker didn\u2019t stop at primary targets,\u201d Dream said. \u201cIt expanded the operation to government IT supply chain vendors, a nuclear safety agency, a government email system, and 7+ energy sector companies \u2014 scanning them all in parallel for misconfigurations, exposed admin interfaces, and exploitable vulnerabilities.\u201d<\/p>\n

It also learned from its mistakes as it went on, Dream said in identifying what stood out about the campaign.<\/p>\n

Autonomous AI-powered cyberattacks have raised alarms in the past: Anthropic reported last fall that it stopped the first autonomous cyber espionage campaign, although researchers noted that the \u201cautonomous\u201d campaign still required significant human work<\/a>.\u00a0<\/p>\n

The Financial Times first reported<\/a> the Dream research and details on the target.<\/p>\n

The hackers used two popular open-source AI frameworks, Hermes and OpenClaw, to set up the Taiwan operation. They bypassed safety guardrails by framing the work as authorized penetration testing, according to Dream.<\/p>\n

The firm discovered the operation via an online archive of 160 megabytes and nearly 1,400 files, revealing \u201ca multi-agent AI system that achieved confirmed, real-world compromises against state infrastructure.\u201d<\/p>\n

As with the autonomous cyber espionage campaign uncovered last fall, the attack Dream examined also noted the need for human tinkering.<\/p>\n

\u201cWe increasingly see threat actors leveraging AI for autonomous offensive operations,\u201d the company wrote. \u201cBut building a system that actually works at this level takes more work than \u2018just\u2019 running a model. It demands careful adjustment to the specific task, optimization of agent coordination, and fine-tuning of decision logic \u2014 the kind of sophistication evident in this framework\u2019s Bayesian prioritization, self-correction loops, and adaptive research cycles.\u201d<\/p>\n

\t\t\t\t\t\"Tim<\/p>\n

\t\t\tWritten by Tim Starks
\n\t\t\tTim Starks is senior reporter at CyberScoop. His previous stops include working at The Washington Post, POLITICO and Congressional Quarterly. An Evansville, Ind. native, he’s covered cybersecurity since 2003. Email Tim here: tim.starks@cyberscoop.com<\/a>.\t\t<\/p>\n","protected":false},"excerpt":{"rendered":"Suspected Chinese hackers used open-source artificial intelligence models to run a cyberattack against the Taiwanese government in the…\n","protected":false},"author":2,"featured_media":138018,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[405,111,7537,387,43199,1590,390],"class_list":["post-138017","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agentic-ai","tag-ai-agents","tag-artificial-intelligence-ai","tag-artificial-intelligence-agents","tag-china","tag-dream","tag-israel","tag-taiwan"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/138017","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=138017"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/138017\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/138018"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=138017"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=138017"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=138017"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}